Ever since I swapped my GeForce 980 ti for a GeForce 1080 Founder's Edition, I have been having a couple of unusual problems that I cannot source.
Frequently, when booting the computer, it will lock up on the motherboard graphic splash (the part where you can hit F11 to get to the boot menu, or del to get to bios). Other times it will hang at a blinking command prompt. If I shut the system down and restart immediately, it will usually boot normally, but sometimes I have to do that more than once. It makes me wince every time.
In addition, I have been having some unusual system crashes that lead to a reboot. I do get the Win10 blue screen with the error message. They don't really seem to be related to anything in the GPU, but who knows? It is not something I ever experienced prior to swapping the cards.
My only idea is that it could somehow be power related? Perhaps the card is exceeding the capability of the PSU? Since swapping the cards, the only thing I have really played is Deus Ex: Mankind Divided, and although I've had some CTDs while playing, the game is pretty much regarded has having a lot of technical issues, so its impossible to tell if the CTDs are due to the game, the card, the system or something in between.
Any ideas?
I can only think of power issues for now.
Jarhead,
I had exactly what you are describing and it was my hard drive failing. It eventually completely failed.
Because you said the problems started after you swapped video cards I would normally say that your problem is related to that. In that case, agreed it could be PSU related or drivers?
But, your symptoms exactly mirror mine when my HD died. If you haven't, I strongly recommend backing up your computer ASAP.
What error messages do you get when you BSOD?
Quote from: Toonces on August 27, 2016, 08:53:29 AM
Jarhead,
I had exactly what you are describing and it was my hard drive failing. It eventually completely failed.
Because you said the problems started after you swapped video cards I would normally say that your problem is related to that. In that case, agreed it could be PSU related or drivers?
But, your symptoms exactly mirror mine when my HD died. If you haven't, I strongly recommend backing up your computer ASAP.
God damn it.
Quote from: Staggerwing on August 27, 2016, 09:04:46 AM
What error messages do you get when you BSOD?
One was a DPC Watchdog Violation. There were a few others though too, that I can't recall.
Did your new GPU have doc with the minimum power rating? If so, how does that compare with your PSU rating?
You haven't had this rig that long have you, JH?
Did you try reseating the card? If you put the old one back in is all well (just wondering if the power supply was 'going').
Quote from: Jarhead0331 on August 27, 2016, 09:30:07 AM
Quote from: Staggerwing on August 27, 2016, 09:04:46 AM
What error messages do you get when you BSOD?
One was a DPC Watchdog Violation. There were a few others though too, that I can't recall.
A quick couple of google chops show that DPC Watchdog Violation errors can be caused by improper shut downs as well as hardware issues (so PSU and GFX card can certainly be the culprit). Did you update your video drivers to the latest version? As DoctorQuest mentioned, verify your GFX card's power needs and if you are just within spec also look at any other major components that are drawing power. Do you have a lot of USB peripherals that need USB power? Do you have a power hungry CPU and/or lots of ginormous cooling fans? All that stuff adds up.
My money is on power
Might be a stupid question, but have you tried swapping GeForce 980 ti card back to see if you still have the issue?
Quick search shows they recommended system power of 500W. Search of card Power consumption is 180 W.
https://www.msi.com/Graphics-card/GTX-1080-Founders-Edition.html
I am an AutoCAD guy and I needed to upgrade to a 750W PSU based on my card, fans, drives. My card needs 142W.
When you restart the computer after a lock, are you hitting reset or powering down and then powering up? I had a computer which would lock every time on a cold boot but a warm boot (reset) at that point would get you through every time as well. This is more of a curiosity question.
As for the instability, I would do the following. Download memtest86 (or whatever the current version is, may have to do some searching) to a usb stick and boot off that. See if the tests complete. If all is good there, start testing from Win10 for stability. RealBench is an excellent testing tool. Make sure you monitor temps while testing though. Get some hd SMART info reader and make sure your OS hd isn't reporting any weirdness.
If something is unstable (which it certainly appears), then the troubleshooting begins. Unless it is obviously a memory or HD issue you could swap back to your old card and retest. If it is the new card, just RMA it. If it is a MB/CPU issue that just cropped up, that is bad news and I would be expecting a failure soon. In the meantime you could monkey with various voltage settings in the BIOS and see if you can get things stable.
As others mentioned, and giving the timing, it could be a PS issue too, though I would think that would be more likely to show up under heavy load.
I would think power or perhaps a heating issue? Have you got any software to detect how hot the system is (and your GPU) when running under load?
Quote from: JudgeDredd on August 27, 2016, 03:55:19 PM
I would think power or perhaps a heating issue? Have you got any software to detect how hot the system is (and your GPU) when running under load?
The issues with boot (appearing at the same time) are pretty strange though. Heating issues shouldn't appear that soon (unless 0 heat sink contact?), though I certainly would check the temp of the cores and GPU as part of the diagnostics and especially during any load testing or GPU bench.
One thing on the GPU side: whatever software you are using, I would setup a user fan profile (%fan speed by degree C), auto can sometimes be wacky. I control my fans and such with MSI afterburner even using an EVGA card, it works fine. I don't have it set to start on boot though (due to overclock settings) and forgot to fire it up the other day and realized that I was sitting at 70C for the entire day (not a huge deal) with StarCrawlers tabbed out. Make sure that whatever you are using for fan control is cooling adequately.
My money is on power as well.
Even though a given PSU might say its a 700w PSU that really doesn't tell you the whole story.
You need a PSU with a constant voltage and amperage output .
Newer PSU's have a dedicated powerline for high end demanding videocards but if you have an older PSU or a second grade brand one, it very possible the thing struggles with providing stable current under load from the GPU.
Quote from: Jarhead0331 on August 27, 2016, 08:34:21 AM
Frequently, when booting the computer, it will lock up on the motherboard graphic splash (the part where you can hit F11 to get to the boot menu, or del to get to bios). Other times it will hang at a blinking command prompt
Quote from: Dread Rlyeh on August 27, 2016, 05:05:47 PM
Quote from: JudgeDredd on August 27, 2016, 03:55:19 PM
I would think power or perhaps a heating issue? Have you got any software to detect how hot the system is (and your GPU) when running under load?
The issues with boot (appearing at the same time) are pretty strange though. Heating issues shouldn't appear that soon (unless 0 heat sink contact?), though I certainly would check the temp of the cores and GPU as part of the diagnostics and especially during any load testing or GPU bench.
Good point
It absolutely sounds like a power supply that isn't able to keep up with your new video card's needs while it is under load.
I've seen that happen more than once, and those are the signs. The BSODs & auto-restarts, especially while running games or movies for a bit, is a good indicator of a hardware-related issue of some sort. Could also be your CPU overheating, while under load, but that usually doesn't happen until you've had the PC for years & stashed away in some location with poor air flow. Could be a few things, but always check power supply first, after doing the driver dance.
Speaking of .. I would check one thing before spending money on a new power supply, if you haven't already:
Uninstall your video drivers, then completely wipe any remaining driver files with a 3rd party driver cleaner program. Then install a fresh & newly downloaded set for your new vid card. Best to be sure there isn't a stray driver file hanging around, still trying to operate your old video card's chipset with your new one.
This is one driver cleaner program you could use: http://www.guru3d.com/files-details/display-driver-uninstaller-download.html
Edit: If you end up needing a new power supply, remember to check the Amperage ratings for the 12v rails on it. The higher the better, obviously. Ya know.. aside from having the appropriate connectors (and #s of them) you need.
Does the new vid card require two separate PCI-E power connectors? Depending on PSU, they may use the same 12v rail thus dividing the current quite a bit. So you'll want a good PSU with high-rated 12v Amps.
in addition to the discussed possible PSU problem
check if there is an update out for your mainboard, its possible that the 1080 is not carefree supported in the PCI E slots you used without a latest patch
you guys are making my head spin. A lot of these proposed fixes are beyond anything I have ever dealt with on my own.
For what it is worth, the 1080 card requires an 8-pin plug from the power supply, but I only had a spare 6-pin and a 2-pin. After doing some quick research I looked to see if a 6+2 combo was acceptable and most sources I came across said it is the same thing. However, could this be part of the problem?
Quote from: Jarhead0331 on August 29, 2016, 03:49:40 PM
you guys are making my head spin. A lot of these proposed fixes are beyond anything I have ever dealt with on my own.
For what it is worth, the 1080 card requires an 8-pin plug from the power supply, but I only had a spare 6-pin and a 2-pin. After doing some quick research I looked to see if a 6+2 combo was acceptable and most sources I came across said it is the same thing. However, could this be part of the problem?
You should know better than ask a bunch of geeks for help, JH. :o
I think at the end of the day the consensus is to swap the old card back in, run for a bit and see if the issue goes away. If it does then at least you've identified the issue as the new card. I've never tried the 6+2 trick myself. It does make me wonder if that might be part of the problem.
Stay frosty (and alert) :) One definitive test you could do would be to swap back in the old card and see if it resolves both the boot and instability issues. Kinda sucks, because given the timing it is probably the card (as in card change - card itself, power supply, bios, etc), but I like to rule stuff out if I can before diving down a rabbit hole that could cost me hours.
If it is for sure the card you could run a GPU bench (like Heaven), and at full draw, if it is a PSU issue, I would expect problems would crop up right quick.
edit: got busy while typing this, DoctorQuest beat me to it.
This sounds very much like a PSU issue. I concur that swapping the old card back in to test is the first thing to try.
I doubt that the split plug is the cause of the issue; as long as you're getting the necessary power it should be fine. The fact that you're seeing the issue at boot only reinforces the idea that this is somehow hardware related.
The other thing to check is BIOS and firmware updates. It is certainly possible that the 1080 requires a newer BIOS or some component (motherboard?) firmware than what you're running. I'd check the OEM websites to see if there are new versions, and if so, update.
It really does feel like a power supply to me, though.
Quote from: Jarhead0331 on August 29, 2016, 03:49:40 PM
you guys are making my head spin. A lot of these proposed fixes are beyond anything I have ever dealt with on my own.
For what it is worth, the 1080 card requires an 8-pin plug from the power supply, but I only had a spare 6-pin and a 2-pin. After doing some quick research I looked to see if a 6+2 combo was acceptable and most sources I came across said it is the same thing. However, could this be part of the problem?
There are 6+2 pin PCI-E connectors and there are 4+4 CPU connectors for some motherboards (AMD iirc).
The 4+4 CPU connector goes directly into the appropriate motherboard plug, whether using only a single 4-pin or both together, as needed.
CPU power looks like this:
(https://www.grogheads.com/forums/proxy.php?request=http%3A%2F%2Fwww.playtool.com%2Fpages%2Fpsuconnectors%2Feps4plus4.jpg&hash=1ee7627969ce68b351ed6ff3abec0780eff03fca)
The PCI-E one plugs into the video card, of course. It
should be labelled on the connector, similar to the following pic, but may not be. However, it should still be a 6+2. Not sure if your specific video card requires TWO connectors plugged in -
some require two separate 6/8 pin connectors.PCI-E (aka Video Card) looks like this:
(https://www.grogheads.com/forums/proxy.php?request=http%3A%2F%2Fwww.mechbgon.com%2Fbuild%2F8-pin_pcie.jpg&hash=98da24dafb1d1793599b27fa979abfae8349d5c6)
Either way, the connectors shouldn't be easily interchangeable since the individual pin molds should be shaped differently between the PCI-E plugs and the CPU plug. Like trying to jam a square shape into a round hole if you are pushing in the wrong one, but I can't guarantee that'll stop someone from doing so.
Probably not related but possibly so I'll add to the thread.
I have a GTX760 that was acquired from a client. He purchased a new system with a GTX760 and he would get these red squares that would appear while in Windows that would then lock up the OS or sometimes when he would boot, the Windows splash screen was screwy or it would just sit at the command prompt. We RMA'd that card and they sent a replacement. THAT card then did the exact same thing. Client was done with the 760 and bought a 9xx series card (and it's working fine). He gave me the 760 (the 2nd one). I had it in one computer and would get the red square issues. If I simply powered down and just pressed on the card, it would more than not boot up and run fine until the next reboot. I've since moved that card to my current rig and that occurs very infrequently but every once in a while when I boot, I get the whacky Windows splash screen so I just power down immediately and press on the card and it almost always boots up normally. A minor inconvenience now (it was much more fussy on my old (old!) rig).
This thread discusses the issue and if you were to dig into Google about "Nvidia red squares" you'll see more posts about it.
http://www.tomshardware.com/answers/id-2175928/gtx-760-red-blue-squares-crash.html
Quote from: Jarhead0331 on August 29, 2016, 03:49:40 PM
you guys are making my head spin. A lot of these proposed fixes are beyond anything I have ever dealt with on my own.
For what it is worth, the 1080 card requires an 8-pin plug from the power supply, but I only had a spare 6-pin and a 2-pin. After doing some quick research I looked to see if a 6+2 combo was acceptable and most sources I came across said it is the same thing. However, could this be part of the problem?
The 6+2 pin combo is fine in itself. I use one on my 1080 as well, but as mentioned above, if the PSU doesn't provide a stable current under demand from the videocard you'll get all kinds of weird issues.
Swap your old card back in and see if the problems persist. If they do not then it's either the 1080 (unlikely) which is faulty or you have a power supply problem.
If the problems do persist with the old card, then you might have damaged the mainboard (static electricity? damaged port?), OR your memory (RAM) is damaged (happens quickly due to static discharge when incorrectly installing new hardware) or there is a timing problem in your BIOS of the mainboard. You can try to reset it to defaults and see if this helps.